CN105490858A - Dynamic link predication method of network structure - Google Patents
Dynamic link predication method of network structure Download PDFInfo
- Publication number
- CN105490858A CN105490858A CN201510929845.3A CN201510929845A CN105490858A CN 105490858 A CN105490858 A CN 105490858A CN 201510929845 A CN201510929845 A CN 201510929845A CN 105490858 A CN105490858 A CN 105490858A
- Authority
- CN
- China
- Prior art keywords
- node
- network configuration
- link
- priority
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/12—Discovery or management of network topologies
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
Abstract
The invention provides a dynamic link predication method of a network structure. The method comprises the following steps: step one, inputting a network structure corresponding to a service object; step two, performing Jaccard distance conversion on the input network structure to obtain a processed network structure; step three, calculating the distance between every two nodes in the network structure; step four, obtaining a network structure with a marked priority at current time; step five, repeatedly executing the step one to the step four at next time to obtain a network structure with a marked priority at the next time, wherein the priority of each link at the next time is postponed to the priority at the current time, and the priorities of the links are successively marked from high to low according to time; and step fix, taking a network structure with a marked priority at each time as a predication result of one network structure for a user to perform analysis processing on the service object. According to the invention, based on a dynamic network topology structure, a dynamic evolution mechanism of a complex network is taken into consideration, the calculation complexity is quite low, and the method provided by the invention is applied to link prediction of a large-scale network.
Description
Technical field
The present invention relates to computer application field, be specifically related to a kind of dynamic link Forecasting Methodology of network configuration.
Background technology
Network system is ubiquity in Human and nature circle.Along with the development of science and technology, people there has also been more deep understanding for the cognition of network system, and the understanding for network system also can help our the cognitive world that we survive further.
Link prediction refers to how by producing the possibility of connection between two nodes not yet producing link in the information prediction networks such as known network configuration.The research of link prediction and the structure of network and evolution contact closely.Current link prediction is emerging, to have a most important theories and using value research direction in complex network research field.
At theoretical side, link prediction can be used for understanding, disclose and possible mechanism of Evolution in more complicated network, by prediction and the comparison of live network, just can obtain the accuracy that this model is predicted network configuration; Link prediction has important application in fields such as influence power analysis, biological scientific experiment, commercial product recommending, Information Communication, network marketings, as biological aspect, in order to avoid blindly detecting the effect of protein by experiment, the interaction of protein first can be estimated by link prediction, thus guiding experiment, avoid the waste of great many of experiments and experiment material.
The method of current link prediction is mainly divided into two large classes, and a class detects based on the similarity of node; Another kind of is detect based on similarity of paths.Existing two large class link prediction methods, mainly based on the network topology structure of static state, can not be taken the dynamic evolution mechanism of complex network into account and have higher computation complexity, not being suitable for the link prediction of large scale network.
Summary of the invention
The invention provides a kind of dynamic link Forecasting Methodology of network configuration, it is based on dynamic network topology structure, takes the dynamic evolution mechanism of complex network into account and to have computation complexity lower, is applicable to the link prediction of large scale network.
The dynamic link Forecasting Methodology of network configuration of the present invention, it comprises the following steps:
Step one, the network configuration that input service object is corresponding;
Step 2, carries out the conversion of Jie Kade distance, obtains the network configuration after process to the network configuration of step one input;
Step 3, utilizes d
(t+1)(u, v)=d
(t)(u, v)+Δ d (u, v) represents the distance in the network configuration after process between any two node u and v; Wherein, d
(t+1)(u, v) represents the distance between t+1 moment u node and v node; d
(t)(u, v) represents the distance between t u node and v node; Distance change in the time period that the t that represents Δ d (u, v) arrives the t+1 moment between u node and v node; And:
Set A is the neighbor node set of v node; Set B is the neighborhood of u node; Deg (u) represents the neighbor node number of u node, or is referred to as the degree of u node; Deg (v) represents the neighbor node number of v node, or is referred to as the degree of v node; Intermediate quantity
with
be expressed as:
λ is the interior poly-coefficient of network configuration;
Step 4, the network configuration of priority-labeled under acquisition current time:
Step 41, calculates the distance in the network configuration after previous moment process between two between node according to the computational methods of step 3;
Step 42, the spacing selecting node between two from the network configuration after previous moment process is all nodes of 0, all performs following process to often pair of node that the spacing of node is between two 0:
By this to node called after u node and v node, and whether there is link in judging the network configuration that u node and v node input in step one, if there is no, then add incoming link between u node and v node in the network configuration of input in step one; If existed, then do not process; Then the priority of this link is marked for highest;
The neighbor node set A of v node in network configuration after the process of traversal previous moment, selecting from neighbor node set A with the spacing of u node is the neighbor node of 0, and by this neighbor node called after x node, and whether there is link in judging the network configuration that u node and x node input in step one, if there is no, then incoming link is added between u node and x node in the network configuration inputted in step one; If existed, then do not process; Then the priority marking this link is second advanced;
The neighbor node set B of u node in network configuration after the process of traversal previous moment, selecting from neighbor node set B with the spacing of v node is the neighbor node of 0, and by this neighbor node called after y node, and whether there is link in judging the network configuration that v node and y node input in step one, if there is no, then incoming link is added between v node and y node in the network configuration inputted in step one; If existed, then do not process; Then the priority marking this link is second advanced;
Wherein, the network configuration of initial time rule of thumb sets;
Step 5, in subsequent time, repeated execution of steps one to step 4, obtains the network configuration of subsequent time priority-labeled, the priority of current time and the priority of each link of this subsequent time postpones, and in chronological sequence marks successively from high to low;
Step 6, the network configuration of each moment priority-labeled is predicting the outcome of a network configuration, and user adds link according to priority.。
The invention has the beneficial effects as follows:
Link prediction method provided by the invention is the angle prediction link from network dynamic evolution mechanism, has taken into account the evolution trend of node interphase interaction and network, has had higher precision of prediction and interpretation.The method can predict the order that link produces, and adds link according to priority orders.Can carry out manual intervention in adding procedure, controllability is stronger; In addition, the method can dynamic analog network evolution process, and computation complexity is low, is applicable to the analysis of large-scale complex network.
Accompanying drawing explanation
Fig. 1 is embodiments of the invention schematic diagram.
Embodiment
Below in conjunction with accompanying drawing, illustrate embodiments of the present invention.
The dynamic link Forecasting Methodology of network configuration of the present invention comprises the following steps:
Step one, the network configuration that input service object is corresponding.
Step 2, utilizes Jie Kade distance (Jaccard) network configuration to initial condition to change, obtains the network configuration after process.
, due to the topology impact between network configuration interior nodes, between two nodes that there is link in the network configuration after process, there is the active force attracted each other in step 3.Utilize d
(t+1)(u, v)=d
(t)(u, v)+Δ d (u, v) represents the distance in the network configuration after process between any two nodes, d
(t+1)(u, v) represents the t+1 moment, the distance between u node and v node; d
(t)(u, v) represents t, the distance between u node and v node; The t that represents Δ d (u, v) arrives between the t+1 moment, the distance change between u node and v node.
Distance between two nodes is constantly change, d
(t+1)(u, v)=d
(t)(u, v)+Δ d
(t)(u, v), Δ d
(t)(u, v) is that the surroundings nodes of link affects the distance of link, and the surroundings nodes of link is the neighbor node of the node at the two ends of link.Δ d
(t)(u, v) can cause shortening or the elongation of euclidean distance between node pair.This process of continuous iteration, distance d (u, v)=0 between u node and v node can be approached to 0 or 1, and when d (u, v)=0 or 1, the distance of link no longer changes.
The impact of surroundings nodes on the distance of link of link has sphere of action, and the interior poly-coefficient lambda be incorporated herein between an expression node carrys out the character of control action power.When the hypertelorism of node, when making interior cluster value be less than the value of λ, now judge that surroundings nodes produces the effect making it extend to the distance of link, when the distance of node is very near, when making interior cluster value be greater than the value of λ, now judge that the distance of surroundings nodes to link has the effect making it shorten.Wherein, λ be test by experiment obtain can the value of reaction network self topological structure, the λ value of heterogeneous networks is different.According to a large amount of tests, λ is value in the scope of (0.2-0.7) generally.
Δ d (u, v) solves shown in following formula:
Wherein, set A is the neighbor node set of v node; Set B is the neighborhood of u node.
Deg (u) represents the neighbor node number of u node, or is referred to as the degree of u node.
Deg (v) represents the neighbor node number of v node, or is referred to as the degree of v node.
Intermediate quantity
solve shown in following formula:
The judgement that link produces.The decision condition that link produces is d (u, v)=0.
Step 4, under current time, if the distance of u and v two nodes becomes 0 i.e. d (u, x)=0, then judge whether there is link in the network configuration that u and v two nodes input in step one, if there is no, then incoming link is added between u and v node in the network configuration inputted in step one; If existed, then do not process; And the priority marking this link is E;
Under current time: for u node, the neighbor node set A of traversal v node, be the situation of 0 to the neighbor node x distance of u node and v node, i.e. d (u, x)=0, then judge whether there is link in the network configuration that u and x two nodes input in step one, if there is no, then add incoming link between u and x node in the network configuration of input in step one; If existed, then do not process; And the priority marking this link is F;
For v node, the neighbor node set B of traversal u node, judge whether that the neighbor node y distance that there is this v node and u node is the situation of 0, i.e. d (v, y)=0, then judge whether there is link in the network configuration that v and y two nodes input in step one, if there is no, then add incoming link between v and y node in the network configuration of input in step one; If existed, then do not process; And the priority marking this link is F;
Above-mentioned process obtains the network configuration of current time afterwards, and this network configuration medium priority E>F;
Step 5, in subsequent time, repeated execution of steps one to step 4, obtain the network configuration medium priority G>H in this moment, wherein, in step one, the network configuration of input is the network configuration obtained in a upper moment, priority E>F>G>H;
Step 6, the network configuration in each moment is one and predicts the outcome, this predict the outcome in the chronological expression forecasting object of link between active force.
As a in Fig. 1 is depicted as the model of initial network, after calling Jie Kade (Jaccard) range formula, obtain the distance between node, reconstructed network model is as shown in b in Fig. 1.Distance is now initial distance and t
0moment distance.Can obtain according to formula, between E, F 2, common node is maximum, and the mutual attractive force be subject to is compared maximum, so subsequent time t
1it is 0 that middle E and F point reaches distance at first, namely forms a set { E, F} (E, F two nodes overlap); Subsequent time t afterwards
2, because the interaction force between B and C point is larger, so subsequent time forms set { B, C}; Proceed distance to solve, t
3d in moment
eH=0; d
fG=0; d
cD=0 in the ideal situation, thinks when H point distance E, F point distance are 0, d
eH=d
fH=0.First judge whether EH, FH and FG, EG have link in reconstructed network, if there is no link, then think and can not add link, if had, then judge whether have link in initial network, then do not add.Then judge whether have link in the reconstructed network between neighbor node G, H, then do not represent and can not occur link in a network.Pass in time thus and carry out network evolution, the network link finally obtained predicts the outcome as shown in c figure in Fig. 1.
Dynamic network link is visual.From initial tense in final steady-state process, along with the change of time, can constantly add new link, when adding incoming link, link is printed to foreground network visualization model part simultaneously, the process that network link adds can be observed dynamically, be convenient to analyze.
Meanwhile, because the process adding incoming link adds according to priority orders, the link that therefore priority is high can highlight, and the display of the link of different priorities also can be had any different, and is convenient to observe.When interpolation link number has requirement, also by priority orders, the link that preferential interpolation priority is high.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, all any amendments done within the spirit and principles in the present invention, equivalent replacement and improvement etc., all should be included within protection scope of the present invention.
Claims (1)
1. a dynamic link Forecasting Methodology for network configuration, is characterized in that, comprise the following steps:
Step one, the network configuration that input service object is corresponding;
Step 2, carries out the conversion of Jie Kade distance, obtains the network configuration after process to the network configuration of step one input;
Step 3, utilizes d
(t+1)(u, v)=d
(t)(u, v)+Δ d (u, v) represents the distance in the network configuration after process between any two node u and v; Wherein, d
(t+1)(u, v) represents the distance between t+1 moment u node and v node; d
(t)(u, v) represents the distance between t u node and v node; Distance change in the time period that the t that represents Δ d (u, v) arrives the t+1 moment between u node and v node; And:
Set A is the neighbor node set of v node; Set B is the neighborhood of u node; Deg (u) represents the neighbor node number of u node, or is referred to as the degree of u node; Deg (v) represents the neighbor node number of v node, or is referred to as the degree of v node; Intermediate quantity
with
be expressed as:
λ is the interior poly-coefficient of network configuration; Wherein, λ be test by experiment obtain can the value of reaction network self topological structure, the λ value of heterogeneous networks is different.According to a large amount of tests, λ is value in the scope of (0.2-0.7) generally.
Step 4, the network configuration of priority-labeled under acquisition current time:
Step 41, calculates the distance in the network configuration after previous moment process between two between node according to the computational methods of step 3;
Step 42, the spacing selecting node between two from the network configuration after previous moment process is all nodes of 0, all performs following process to often pair of node that the spacing of node is between two 0:
By this to node called after u node and v node, and whether there is link in judging the network configuration that u node and v node input in step one, if there is no, then add incoming link between u node and v node in the network configuration of input in step one; If existed, then do not process; Then the priority of this link is marked for highest;
The neighbor node set A of v node in network configuration after the process of traversal previous moment, selecting from neighbor node set A with the spacing of u node is the neighbor node of 0, and by this neighbor node called after x node, and whether there is link in judging the network configuration that u node and x node input in step one, if there is no, then incoming link is added between u node and x node in the network configuration inputted in step one; If existed, then do not process; Then the priority marking this link is second advanced;
The neighbor node set B of u node in network configuration after the process of traversal previous moment, selecting from neighbor node set B with the spacing of v node is the neighbor node of 0, and by this neighbor node called after y node, and whether there is link in judging the network configuration that v node and y node input in step one, if there is no, then incoming link is added between v node and y node in the network configuration inputted in step one; If existed, then do not process; Then the priority marking this link is second advanced;
Step 5, in subsequent time, repeated execution of steps one to step 4, obtains the network configuration of subsequent time priority-labeled, the priority of current time and the priority of each link of this subsequent time postpones, and in chronological sequence marks successively from high to low;
Step 6, the network configuration of each moment priority-labeled is predicting the outcome of a network configuration, and priority provides reference for user adds link.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510929845.3A CN105490858B (en) | 2015-12-15 | 2015-12-15 | A kind of dynamic link prediction technique of network structure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510929845.3A CN105490858B (en) | 2015-12-15 | 2015-12-15 | A kind of dynamic link prediction technique of network structure |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105490858A true CN105490858A (en) | 2016-04-13 |
| CN105490858B CN105490858B (en) | 2018-08-03 |
Family
ID=55677601
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510929845.3A Expired - Fee Related CN105490858B (en) | 2015-12-15 | 2015-12-15 | A kind of dynamic link prediction technique of network structure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105490858B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106559290A (en) * | 2016-11-29 | 2017-04-05 | 北京邮电大学 | Method and system based on the link prediction of community structure |
| CN107623586A (en) * | 2017-07-20 | 2018-01-23 | 深圳大学 | Network link Forecasting Methodology and device |
| CN107958289A (en) * | 2016-10-18 | 2018-04-24 | 深圳光启合众科技有限公司 | Data processing method and device, robot for robot |
| CN110276113A (en) * | 2019-06-11 | 2019-09-24 | 嘉兴深拓科技有限公司 | A kind of network structure prediction technique |
| CN110442751A (en) * | 2019-06-27 | 2019-11-12 | 浙江工业大学 | Dynamic link prediction meanss and application based on production confrontation network |
| CN111881977A (en) * | 2020-07-27 | 2020-11-03 | 复旦大学 | Community network detection method, device and system based on inter-node distance trend |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2003090122A2 (en) * | 2002-04-19 | 2003-10-30 | Computer Associates Think, Inc. | Using neural networks for data mining |
| DE10324045B3 (en) * | 2003-05-27 | 2004-10-14 | Siemens Ag | System characteristics modelling method for dynamic system using similarity analysis for modification of known system characteristics supplied to neural network structure for causality analysis |
| CN104899657A (en) * | 2015-06-09 | 2015-09-09 | 北京邮电大学 | Method for predicting association fusion events |
-
2015
- 2015-12-15 CN CN201510929845.3A patent/CN105490858B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2003090122A2 (en) * | 2002-04-19 | 2003-10-30 | Computer Associates Think, Inc. | Using neural networks for data mining |
| DE10324045B3 (en) * | 2003-05-27 | 2004-10-14 | Siemens Ag | System characteristics modelling method for dynamic system using similarity analysis for modification of known system characteristics supplied to neural network structure for causality analysis |
| CN104899657A (en) * | 2015-06-09 | 2015-09-09 | 北京邮电大学 | Method for predicting association fusion events |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107958289A (en) * | 2016-10-18 | 2018-04-24 | 深圳光启合众科技有限公司 | Data processing method and device, robot for robot |
| CN106559290A (en) * | 2016-11-29 | 2017-04-05 | 北京邮电大学 | Method and system based on the link prediction of community structure |
| CN106559290B (en) * | 2016-11-29 | 2019-09-27 | 北京邮电大学 | The method and system of link prediction based on community structure |
| CN107623586A (en) * | 2017-07-20 | 2018-01-23 | 深圳大学 | Network link Forecasting Methodology and device |
| CN107623586B (en) * | 2017-07-20 | 2021-06-04 | 深圳大学 | Network link prediction method and device |
| CN110276113A (en) * | 2019-06-11 | 2019-09-24 | 嘉兴深拓科技有限公司 | A kind of network structure prediction technique |
| CN110442751A (en) * | 2019-06-27 | 2019-11-12 | 浙江工业大学 | Dynamic link prediction meanss and application based on production confrontation network |
| CN110442751B (en) * | 2019-06-27 | 2022-04-29 | 浙江工业大学 | Dynamic link prediction device based on generative countermeasure network and application |
| CN111881977A (en) * | 2020-07-27 | 2020-11-03 | 复旦大学 | Community network detection method, device and system based on inter-node distance trend |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105490858B (en) | 2018-08-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105490858A (en) | Dynamic link predication method of network structure | |
| Wang et al. | A novel Elite Opposition-based Jaya algorithm for parameter estimation of photovoltaic cell models | |
| Sayed | Adaptive networks | |
| Zhou et al. | Invasive weed optimization algorithm for optimization no-idle flow shop scheduling problem | |
| Zhao et al. | Self-adaptive differential evolution with multi-trajectory search for large-scale optimization | |
| Zhang et al. | Modeling and monitoring between-mode transition of multimodes processes | |
| Wu et al. | Modeling a SOFC stack based on GA-RBF neural networks identification | |
| Li et al. | Improved gene expression programming to solve the inverse problem for ordinary differential equations | |
| Marappan et al. | A new genetic algorithm for graph coloring | |
| Huang et al. | Electromechanical equipment state forecasting based on genetic algorithm–support vector regression | |
| Apolloni et al. | Empirical evaluation of distributed differential evolution on standard benchmarks | |
| CN106296044A (en) | power system risk scheduling method and system | |
| Aliyu et al. | Efficient metaheuristic population-based and deterministic algorithm for resource provisioning using ant colony optimization and spanning tree | |
| Kamelian et al. | A novel graph-based partitioning algorithm for large-scale dynamical systems | |
| Kapilevich et al. | Chromatin 3D reconstruction from chromosomal contacts using a genetic algorithm | |
| Karami | A framework for uncertainty-aware visual analytics in big data | |
| Li et al. | Minimum energy control and optimal-satisfactory control of Boolean control network | |
| Manoj et al. | Multi-objective optimization through a novel Bayesian approach for industrial manufacturing of Polyvinyl Acetate | |
| Lingjærde et al. | Scalable multiple network inference with the joint graphical horseshoe | |
| Sone et al. | Forecasting wireless network traffic and channel utilization using real network/physical layer data | |
| Gao et al. | Application of BPSO with GA in model-based fault diagnosis of traction substation | |
| Wang et al. | Power system transient security assessment based on multi-channel time series data mining | |
| Muros et al. | Cooperative game theory tools to detect critical nodes in distributed control systems | |
| Kässinger et al. | Persival: Simulating complex 3d meshes on resource-constrained mobile ar devices using interpolation | |
| Snyder et al. | Design and modeling for self-organizing autonomic systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180803 Termination date: 20191215 |